Re: [ovs-dev] [PATCH 5/7] utils: Introduce xusleep for subsecond granularity.

[Aaron Conole]

"Bodireddy, Bhanuprakash"  writes:

>>Bhanuprakash Bodireddy  writes:
>>
>>> This will be used by KA framework that needs millisecond granularity.
>>>
>>> Signed-off-by: Bhanuprakash Bodireddy
>>> 
>>> ---
>>
>>Without this patch, builds starting at 3/7 will fail.  That's pretty
>> bad.  Please see
>>my earlier comment about including helpers when they are used.
>
> As mentioned earlier, it was my mistake and will fix this in v2.
>
>>
>>>  lib/util.c | 12 
>>>  lib/util.h |  1 +
>>>  2 files changed, 13 insertions(+)
>>>
>>> diff --git a/lib/util.c b/lib/util.c
>>> index 1c06ce0..889ebd8 100644
>>> --- a/lib/util.c
>>> +++ b/lib/util.c
>>> @@ -2125,6 +2125,18 @@ xsleep(unsigned int seconds)
>>>  ovsrcu_quiesce_end();
>>>  }
>>>
>>> +void
>>> +xusleep(unsigned int microseconds)
>>> +{
>>> +ovsrcu_quiesce_start();
>>> +#ifdef _WIN32
>>> +Sleep(microseconds/1000);
>>> +#else
>>> +usleep(microseconds);
>>> +#endif
>>> +ovsrcu_quiesce_end();
>>> +}
>>> +
>>
>>Wow!  This is deceptive.  If I call this with microseconds argument
>> as, say, 999
>>there's a *strong* chance this will NOT sleep for at least that amount of 
>>time.
>>This function needs a different implementation or just keep it non-windows.
>
> This is indeed deceptive :).  
> I did spend some time to understand if there is an equivalent usleep()
> implementation in Windows and came across Queryperformancecounter and
> usermode scheduling in windows.  I would better stick to non-windows
> implementation and allow the WINDOWS experts add their equivalent
> implantation above.

Since DPDK is only for linux/BSD right now, better to not expose this.
Let it be implemented when it's needed.  I'd make it a static inside
the dpdk.c file.  If the functionality is required from the windows
side, it can be re-abstracted.

> - Bhanuprakash.
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

Re: [ovs-dev] [PATCH 5/7] utils: Introduce xusleep for subsecond granularity.

[Bodireddy, Bhanuprakash]

>Bhanuprakash Bodireddy  writes:
>
>> This will be used by KA framework that needs millisecond granularity.
>>
>> Signed-off-by: Bhanuprakash Bodireddy
>> 
>> ---
>
>Without this patch, builds starting at 3/7 will fail.  That's pretty bad.  
>Please see
>my earlier comment about including helpers when they are used.

As mentioned earlier, it was my mistake and will fix this in v2.

>
>>  lib/util.c | 12 
>>  lib/util.h |  1 +
>>  2 files changed, 13 insertions(+)
>>
>> diff --git a/lib/util.c b/lib/util.c
>> index 1c06ce0..889ebd8 100644
>> --- a/lib/util.c
>> +++ b/lib/util.c
>> @@ -2125,6 +2125,18 @@ xsleep(unsigned int seconds)
>>  ovsrcu_quiesce_end();
>>  }
>>
>> +void
>> +xusleep(unsigned int microseconds)
>> +{
>> +ovsrcu_quiesce_start();
>> +#ifdef _WIN32
>> +Sleep(microseconds/1000);
>> +#else
>> +usleep(microseconds);
>> +#endif
>> +ovsrcu_quiesce_end();
>> +}
>> +
>
>Wow!  This is deceptive.  If I call this with microseconds argument as, say, 
>999
>there's a *strong* chance this will NOT sleep for at least that amount of time.
>This function needs a different implementation or just keep it non-windows.

This is indeed deceptive :).  
I did spend some time to understand if there is  an equivalent usleep() 
implementation in Windows and came across Queryperformancecounter and usermode 
scheduling in windows.  I would better stick to non-windows implementation and 
allow the WINDOWS experts add their equivalent implantation above. 

- Bhanuprakash.
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

Re: [ovs-dev] [PATCH 5/7] utils: Introduce xusleep for subsecond granularity.

[Aaron Conole]

Bhanuprakash Bodireddy  writes:

> This will be used by KA framework that needs millisecond granularity.
>
> Signed-off-by: Bhanuprakash Bodireddy 
> ---

Without this patch, builds starting at 3/7 will fail.  That's pretty
bad.  Please see my earlier comment about including helpers when they
are used.

>  lib/util.c | 12 
>  lib/util.h |  1 +
>  2 files changed, 13 insertions(+)
>
> diff --git a/lib/util.c b/lib/util.c
> index 1c06ce0..889ebd8 100644
> --- a/lib/util.c
> +++ b/lib/util.c
> @@ -2125,6 +2125,18 @@ xsleep(unsigned int seconds)
>  ovsrcu_quiesce_end();
>  }
>  
> +void
> +xusleep(unsigned int microseconds)
> +{
> +ovsrcu_quiesce_start();
> +#ifdef _WIN32
> +Sleep(microseconds/1000);
> +#else
> +usleep(microseconds);
> +#endif
> +ovsrcu_quiesce_end();
> +}
> +

Wow!  This is deceptive.  If I call this with microseconds argument as,
say, 999 there's a *strong* chance this will NOT sleep for at least that
amount of time.  This function needs a different implementation or just
keep it non-windows.

>  /* Determine whether standard output is a tty or not. This is useful to 
> decide
>   * whether to use color output or not when --color option for utilities is 
> set
>   * to `auto`.
> diff --git a/lib/util.h b/lib/util.h
> index aa38122..637d0c3 100644
> --- a/lib/util.h
> +++ b/lib/util.h
> @@ -451,6 +451,7 @@ ovs_u128_and(const ovs_u128 a, const ovs_u128 b)
>  }
>  
>  void xsleep(unsigned int seconds);
> +void xusleep(unsigned int microseconds);
>  
>  bool is_stdout_a_tty(void);
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

[ovs-dev] [PATCH 5/7] utils: Introduce xusleep for subsecond granularity.

[Bhanuprakash Bodireddy]

This will be used by KA framework that needs millisecond granularity.

Signed-off-by: Bhanuprakash Bodireddy 
---
 lib/util.c | 12 
 lib/util.h |  1 +
 2 files changed, 13 insertions(+)

diff --git a/lib/util.c b/lib/util.c
index 1c06ce0..889ebd8 100644
--- a/lib/util.c
+++ b/lib/util.c
@@ -2125,6 +2125,18 @@ xsleep(unsigned int seconds)
 ovsrcu_quiesce_end();
 }
 
+void
+xusleep(unsigned int microseconds)
+{
+ovsrcu_quiesce_start();
+#ifdef _WIN32
+Sleep(microseconds/1000);
+#else
+usleep(microseconds);
+#endif
+ovsrcu_quiesce_end();
+}
+
 /* Determine whether standard output is a tty or not. This is useful to decide
  * whether to use color output or not when --color option for utilities is set
  * to `auto`.
diff --git a/lib/util.h b/lib/util.h
index aa38122..637d0c3 100644
--- a/lib/util.h
+++ b/lib/util.h
@@ -451,6 +451,7 @@ ovs_u128_and(const ovs_u128 a, const ovs_u128 b)
 }
 
 void xsleep(unsigned int seconds);
+void xusleep(unsigned int microseconds);
 
 bool is_stdout_a_tty(void);
 
-- 
2.4.11

___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev